Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/jiahaoxiang2000/lbox-search
https://github.com/jiahaoxiang2000/lbox-search
Last synced: about 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/jiahaoxiang2000/lbox-search
- Owner: jiahaoxiang2000
- Created: 2024-10-06T08:05:01.000Z (3 months ago)
- Default Branch: main
- Last Pushed: 2024-11-07T00:18:23.000Z (2 months ago)
- Last Synced: 2024-11-07T01:24:54.727Z (2 months ago)
- Language: C
- Size: 19.5 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Design of a Linear Layer Optimised for Bitsliced 32-bit Implementation
> NOTE: this repository is the copy of the original repository. The original repository is available at [github origin](https://github.com/Lbox-ToSC/lbox-search).
This repository contains additional data and code from the paper:
> [*Design of a Linear Layer Optimised for Bitsliced 32-bit Implementation*](https://tosc.iacr.org/index.php/ToSC/article/view/11412)
> Gaëtan Leurent and Clara Pernot
> ToSC 2024(1)
## Repository content
This repository includes:
1. Code to search for efficient linear transforms, in file
`find_lbox.c`;
2. Code to search for a relatively efficient implementation of a linear
transform, in file `heuristic_implem.c`;
3. Reference implementation of the two linear transforms described in
the paper, in files `L32x3.c` and `L32x4.c`.
## 1. Finding efficient linear transforms: `find_lbox.c`
This program implements the main search of the paper.
There are three ways t use the program
1. Look for random circulant matrices with high branch number
```
./find_lbox -r
```
2. Look for circulant matrices with efficient implementation and high branch number
```
./find_lbox
```
3. Verify the branch number of a given LBox (the example corresponds to L32x4 from the paper)
```
./find_lbox 44032028488021000802200841114541
```
Important parameters are `#define`ed at the top of the file:
- `WORD_SIZE`: parameter $\ell$ in the paper
- `WORD_NB`: parameter $w$ in the paper
- `TARGET_WEIGHT`: maximum weight for the search (inclusive). In order to proove that the branch number is at least w, set `TARGET_WEIGHT` to w-1 in order to verify that no words of weight w or less are found.
- `ITER`: number of iterations (log2)
- `MAX_XOR`: number of rotation+XOR steps in the implementation
- `COMBINE3ROWS`: use combinaisons of up to three rows ($d=3$ in the paper). Otherwise, use combinaisons of up to two rows ($d=2$)
Some parameter examples follow
### Parameters for an LBox over 3 32-bit words
```
#define WORD_SIZE 32
#define WORD_NB 3
#define TARGET_WEIGHT 18
#define ITER 10
#define MAX_XOR 6
#define COMBINE3ROWS
```
### Parameters for an LBox over 4 32-bit words
```
#define WORD_SIZE 32
#define WORD_NB 4
#define TARGET_WEIGHT 20
#define ITER 10
#define MAX_XOR 6
```
### Parameters for an LBox over 4 16-bit words
```
#define WORD_SIZE 16
#define WORD_NB 4
#define TARGET_WEIGHT 11
#define ITER 10
#define MAX_XOR 5
```
### Parameters for an LBox over 8 16-bit words
```
#define WORD_SIZE 16
#define WORD_NB 4
#define TARGET_WEIGHT 13
#define ITER 12
#define MAX_XOR 6
```
## 2. Finding an implementation of a linear transforms: `heuristic_implem.c`
This code uses the `_BitInt` feature of C23 for variable-length
integers. Therefore you need a recent version of `clang` (>=15) in
order to compile it.
The word size is assumed to bit 32 ($\ell = 32$).
The `WIDTH` parameter defined at the top of the file corresponds to the parameter $w$ in the paper
Example (the example corresponds to the inverse of L32x4 from the paper):
```
./heuristic_implem 81034926c0d0e290b0606d93f2049851
```